Comparison of fit for a future purpose concept based on tolerance interval and ISO 11352 approaches to predict the quantitative performances and routine uncertainty of an ICP-MS method for the determination of 21 elements in drinking water

This article outlines the validation results of an analytical method for the dosage at the same time twenty-one trace and ultra-trace elements- Li, Be, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Mo, Cd, Sn, Sb, Pb, Ag and Hg (using gold as a mercury stabilizer) - by ICP-MS in drinking water samples. A new validation methodology based on the expanded uncertainty profiles was exploited to select the best suitable calibration model, to assess the method accuracy within well-defined acceptability limits and to determine the limits of quantification as well as the working range. By means of an innovative formula based on the uncertainty concept, the validation study was completed and supported by the investigation of the risk profiles of various acceptable regression models with the aim of minimizing the risk of either accepting an invalid method or refusing a valid method. On the other hand, a new way of calculating the uncertainty of analytical results derived from beta-content gamma-confidence, tolerance interval as a complement to the top-down approaches proposed to date was described. The uncertainty estimates obtained by this proposal from validation data were compared via F-test with those arising from the ISO-11352 norm. In the light of this comparison, it can be concluded that our strategy makes it possible to approximate perfectly the routine uncertainty. (C) 2018 Elsevier B.V. All rights reserved.